Dolly
Family: GPT
Pretraining Architecture: Decoder
Fine-tuning Task: human instructions
Extension: fine-tuned based on the GPT-J-6B (V1) and Pythia model (V2)
Application: Similar to Alpaca
Date (of first known publication): 03/2023
Num. Params: V1: 6B, V2: 12B
Corpus: V1: Instruction corpus same as Alpaca, V2: databricks own dataset.
License: Open
Lab: Databricks, Inc
Last updated